First post! Hehe 😄

I'm testing offprint.app. It's not a classic long-form blog engine or platform. In WordPress or other legacy solutions the architecture is pretty straightforward and also limited: you interact with a web application (written in PHP ages ago, if it's WordPress) and your content is stored in databases such as MySQL somewhere in the cloud (you can self-host it, use a professionally administered paid service, etc.) It's all monolithic and centralized.

In the old world, people can use RSS or Atom (or WordPress.com notifications via app or email) to follow you and get notified about your new posts. If you want to reach friends who don't follow your blog you have to share the URL of the new post in different social networks. This can be automated, but as you can see everything is pretty ad-hoc and disconnected. You can add plug-ins to WordPress to share content to ActivityPub (used by Mastodon) and probably also the ATmosphere, but again, it's not a native feature, it's an add-on.

This is different. Bluesky uses the new AT Protocol which is way more modern, and federated. You have an account with a persistent identity (DID = decentralized identifier) and a mutable handle which is usually a subdomain, but you can use a custom domain (I'm using @imouthes.com). If you own it, you just have to prove it by adding a TXT DNS record (or publishing a token in a well-known URL).

Your identity, content, etc. resides in a PDS (= personal data server). Through single sign-on (OAuth) you allow different applications (web apps, mobile apps, etc.) to access your records. The main one is the reference implementation, Bluesky, which looks a bit like the old Twitter. But there are many others. People have already implemented end-to-end encrypted messaging such as Germ DM, and long-form blogging engines such as the one I'm testing right now, Offprint.

The technical details of the protocol are interesting. It stores your data in records, with strongly-typed JSON open schemas (Lexicons). Offprint and other AT Protocol publishing tools use standard.site. There is something called the firehose, relays and indexers. When you publish an article, the PDS emits an event to the global firehose (I'm trying to write a simple explanation here). Applications listening to the firehose, including search engines, can know about it without having to crawl a web site the old way.

There is no strong "vendor lock-in", because the format is open and standard, the Offprint app is just an editor and viewer of my posts. Migration is structurally easier. And the social "backbone" is native. Replies on Bluesky can sync natively with the blog, etc. Also, instead of fragile HTTP(S) URLs that can change and stop working with infrastructure changes, posts have a persistent AT-URI, something like this: at://did:plc:[your-unique-id]/site.standard.document/[record-key].

This is not a distributed peer-to-peer filesystem such as IPFS, it's more like a federated, distributed social platform, with open standards for data and applications. We'll see how it goes, so far it's fun just testing it.